-
Notifications
You must be signed in to change notification settings - Fork 115
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
"SimpleKernel" example throws exception #36
Comments
hi @AlexBatsev, have you tried updating your CUDA drivers? This appears to be a similar problem to #29. |
After I've updated drivers the problem's gone. |
It makes sense to add additional driver version checks to check compatibility. We should add them as soon as possible to improve usability :)
In general: no. We had to update the PTX version to support new GPU features. However, the upcoming release of the ILGPU compiler has a completely new intrinsic processing pipeline. This allows us to define version-dependent implementations and specialized code generation for different GPU targets. This should help us to work around the problem of general GPU support for older hardware platforms that you mentioned. |
@m4rs-mt thank you for answers. |
- Added support for Half type. - Added O2 configuration to unit tests. - Restrict algorithms to 32-bit ArrayView allocations.
I've tried to run SimpleKernel example and has got
ILGPU.Runtime.Cuda.CudaException: 'a PTX JIT compilation failed'
Inside CudaKernel constructor, the function
CudaAPI.Current.LoadModule(...)
returnedCudaError cudaStatus = CUDA_ERROR_INVALID_PTX;
Indeed, on line 5 one can see ".version 6.0":
kernel.PTXAssembly =
If it is important, my graphic card is Quadro M2000M.
Is this a bug or do I misunderstand something?
Thank you.
The text was updated successfully, but these errors were encountered: